Journal of Molecular Biology
Volume 425, Issue 2, 23 January 2013, Pages 350-364
Journal home page for Journal of Molecular Biology

A 2.1-Å-Resolution Crystal Structure of Unliganded CRM1 Reveals the Mechanism of Autoinhibition

https://doi.org/10.1016/j.jmb.2012.11.014Get rights and content

Abstract

CRM1 mediates nuclear export of numerous proteins and ribonucleoproteins containing a leucine-rich nuclear export signal (NES). Binding of RanGTP to CRM1 in the nucleus stabilizes cargo association with CRM1, and vice versa, but the mechanism underlying the positive cooperativity in RanGTP and NES binding to CRM1 remains incompletely understood. Herein we report a 2.1-Å-resolution crystal structure of unliganded Saccharomyces cerevisiae CRM1 (Xpo1p) that demonstrates that an internal loop of CRM1 (referred to as HEAT9 loop) is primarily responsible for maintaining the NES-binding cleft in a closed conformation, rendering CRM1 incapable of NES binding in the absence of RanGTP. The structure also shows that the C-terminal tail of CRM1 stabilizes the autoinhibitory conformation of the HEAT9 loop and thereby reinforces autoinhibition. Comparison with the structures of CRM1–NES–RanGTP complexes reveals how binding of RanGTP is associated with a series of allosteric conformational changes in CRM1 that lead to opening of the NES-binding cleft, allowing for stable binding of NES cargoes.

Graphical Abstract

Highlights

► Mechanism of cooperative RanGTP and cargo binding to CRM1 has remained obscure. ► We report a 2.1-Å-resolution crystal structure of unliganded S. cerevisiae CRM1. ► The HEAT9 loop is primarily responsible for autoinhibition of CRM1. ► The C-terminus of CRM1 reinforces autoinhibition by the HEAT9 loop. ► The structure advances understanding of allosteric coupling in CRM1.

Introduction

Macromolecular exchange between the nucleus and the cytoplasm is an essential cellular process in all eukaryotes and occurs through the nuclear pore complexes (NPCs) embedded in the nuclear envelope. Most transport events through the NPCs are mediated by multiple families of soluble transport receptors.[1], [2] The cargo macromolecules bind to specific transport receptors in either the cytoplasm or the nucleus and are then translocated through NPCs, after which the transport receptors release their cargoes and are recycled to the original compartment to participate in another transport cycle. The largest class of the nuclear transport receptors is the family of importin-β-like transport factors designated as karyopherins, which can be classified into two types, importins and exportins, depending on the directionality of transport. Importins carry cargoes to the nucleus, whereas exportins carry cargoes to the cytoplasm.

The small GTPase Ran regulates karyopherin–cargo interactions and the directionality of karyopherin-mediated transport. Like other Ras-family GTPases, Ran cycles between GTP- and GDP-bound states, and the two surface loops in Ran, referred to as the switch I and switch II loops, undergo significant conformational changes between the GTP- and GDP-bound states.3 In addition, Ran has a C-terminal extension that is disordered in the GTP-bound state but folds back against the body of Ran as an α-helix in the GDP-bound state.3 The conformations of these three regions (switch I, switch II, and the C-terminal extension) that are sensitive to the nucleotide state of Ran are important in determining its interactions with other proteins.4 The low intrinsic rates of nucleotide exchange and hydrolysis on Ran are stimulated by specific factors in vivo. Ran GTPase activity is stimulated by the order of 105 by RanGAP,5 whereas nucleotide exchange is stimulated by the Ran guanine nucleotide exchange factor, RCC1.6 Based on the localizations of RanGAP in the cytoplasm and RCC1 in the nucleoplasm, cytoplasmic Ran is primarily in the GDP-bound state whereas nucleoplasmic Ran is kept primarily in the GTP-bound state. This gradient of Ran nucleotide state is an important determinant of the directionality of nuclear transport.7 In nuclear import, RanGTP competes with the cargoes to bind to importins, allowing the cargo binding in the cytoplasm and RanGTP-mediated cargo dissociation in the nucleus. By contrast, in nuclear export, RanGTP and cargoes bind cooperatively to exportins in the nucleus, and the exportin–RanGTP–cargo complexes are disassembled in the cytoplasm, where RanGTPase is activated by RanGAP and the Ran-binding proteins RanBP1/2. Thus, the association and dissociation of karyopherin–cargo complexes are regulated by direct binding of Ran in a compartment-specific manner.

CRM1 (also known as exportin 1 or Xpo1) is the most versatile exportin that facilitates nuclear export of a broad range of cargoes.[8], [9], [10], [11] The majority of the export substrates of CRM1 contain a short peptide sequence (10–15 residues), the so-called leucine-rich nuclear export signal (NES) that was first identified in cAMP-dependent protein kinase inhibitor (PKI)12 and HIV-1 Rev.13 Leucine-rich NESs typically harbor four or five characteristically spaced hydrophobic residues that are crucial for the binding to CRM1. Although the NESs show considerable sequence diversity,[14], [15] recent structure determination of CRM1–cargo complexes with and without RanGTP[16], [17], [18] showed that at least three NESs (leucine-rich NES of snurportin, PKI, and Rev) share the ability to bind specifically to the same site: a hydrophobic cleft of CRM1. CRM1 is a ring-shaped molecule that is constructed from 21 tandem HEAT repeats, each of which consists of two antiparallel α-helices, designated A-helix and B-helix, connected by loops of varying length. The A-helices form outer convex surface whereas the B-helices form the inner concave surface. The hydrophobic cleft on the convex outer surface of CRM1, formed between the A-helices of HEAT repeats 11 and 12, constitutes the NES-binding site. The hydrophobic side chains of NESs fit into five hydrophobic pockets along this cleft. The binding site of Leptomycin B, a potent inhibitor of CRM1-mediated nuclear export,9 is located in this hydrophobic cleft,19 and so it is likely that this cleft is the general binding site for NESs. CRM1 is unusual among karyopherins in that it has a cargo-binding site on its outer surface (instead of its inner surface),[4], [20], [21], [22], [23] but this is important for CRM1 to carry a broad range of cargoes that vary greatly in size and shape, including huge cargoes such as ribosomal subunits. In contrast to the cargoes that bind to the outer surface of CRM1, RanGTP binds to the inner surface of CRM1.[17], [18] Four distinct regions of CRM1 contribute to RanGTP binding, and CRM1 directly binds to both switch I and switch II loops of RanGTP, accounting for the ability of CRM1 to discriminate between GTP- and GDP-bound Ran. Interestingly, the C-terminal α-helix of CRM1 (the C-helix) adopts dramatically different positions depending on whether or not RanGTP is bound to the CRM1–cargo complexes.[16], [17] In the binary CRM1–snurportin complex, the C-helix lies across the central cavity of the CRM1 ring with its C-terminus located close to the NES-binding site, whereas in the ternary CRM1–snurportin–RanGTP complex, the C-helix is located on the outer surface of the CRM1 ring, indicating that the C-terminus of CRM1 could regulate the affinity of NES in a way that is sensitive to RanGTP.

In the cytoplasm, the ternary CRM1–cargo–RanGTP complex is disassembled by the action of Ran-binding proteins RanBP1/2 and RanGAP. A recent kinetic study showed that the Ran-binding domains (RanBDs) of the cytoplasmic proteins RanBP1 and RanBP2, but not RanGAP, accelerate dissociation of NES from CRM1 and RanGTP by over 2 orders of magnitude.24 Crystal structure of yeast CRM1–RanBP1–RanGTP complex showed that the NES- and RanBD-binding sites on CRM1 are distinct and that the binding of RanBD induces the movement of a long internal loop in HEAT repeat 9 (referred to as HEAT9 loop), from switch I of RanGTP to the concave side of the NES-binding cleft (the inner surface of HEAT repeats 11 and 12 of CRM1), driving rotations and translations of the α-helices constituting the NES-binding site. This results in closure of the hydrophobic cleft to dissociate NES.24 Structure-based mutagenesis of crucial hydrophobic residues of the HEAT9 loop provided strong support for this allosteric mechanism of NES release and also indicated that the HEAT9 loop functions as an allosteric autoinhibitor to stabilize CRM1 in a conformation that is unable to bind NES cargo in the absence of RanGTP.24 Mutational analyses also indicated that the C-terminus of CRM1 plays an important role in stabilizing CRM1 in an autoinhibited state in the absence of RanGTP,[25], [26] and so it has been proposed that the HEAT9 loop and the C-terminus of CRM1 cooperate to stabilize the NES-binding cleft in a closed state in the absence of RanGTP.[24], [26] Nevertheless, the absence of a high-resolution structure of CRM1 in isolation meant that the structural basis for autoinhibition remained obscure.

Here, we describe a 2.1-Å-resolution crystal structure of unliganded CRM1 that provides definitive structural data to establish the precise mechanism of autoinhibition of CRM1. This structure provides direct evidence that the NES-binding cleft of the unliganded CRM1 adopts a closed conformation, which is stabilized by the binding of the HEAT9 loop to the inner surface of HEAT repeats 11 and 12 in exactly the same way as observed in the CRM1–RanBP1–RanGTP complex. Moreover, our new structure shows that the C-terminus interacts with the inner surface of CRM1 beneath the NES-binding site in such a way that stabilizes the HEAT9 loop in the autoinhibitory conformation. Thus, the structure reported here highlights the crucial role of the HEAT9 loop as the primary determinant of the conformation of the NES-binding cleft and also reveals how the autoinhibitory function of the HEAT9 loop is reinforced by the C-terminus. Comparison with the structure of CRM1–cargo–RanGTP complexes reveals how binding of RanGTP induces movement of both the HEAT9 loop and the C-terminus of CRM1, enabling the transition of the NES-binding cleft from the closed state to the open state.

Section snippets

Crystallization and structure determination of unliganded CRM1

To understand the structural basis for autoinhibition of CRM1, we attempted to crystallize full-length CRM1 from various species including Homo sapiens, Saccharomyces cerevisiae, and Schizosaccharomyces pombe. Although extensive attempts to crystallize full-length, wild-type CRM1 were unsuccessful, optimization of the S. cerevisiae CRM1 (Xpo1p) construct yielded diffraction-quality crystals suitable for high-resolution structure determination. We employed a strategy to delete unnecessary

Expression and purification of yCRM1 for crystallization

Glutathione S-transferase (GST)-CRM1 (S. cerevisiae, Xpo1p, a deletion mutant in which residues 377–413 and residues 971–984 are deleted) was expressed from pGEX-tobacco etch virus (TEV)29 in the Escherichia coli host strain BL21-CodonPlus(DE3)RIL (Stratagene) in 2 × TY medium at 20 °C. After harvesting, the pellet was frozen in liquid nitrogen and stored at − 20 °C until needed.

For purification, the frozen pellet of the cells expressing GST-CRM1 were thawed at room temperature and resuspended in

Acknowledgements

We thank our colleagues in Nagoya, especially Masako Koyama, Junya Kobayashi, Hidemi Hirano, and Tatsuo Hikage, for assistance and discussion. We are also indebted to the staff of Photon Factory and SPring-8 for assistance during data collection. This work was supported in part by the Sumitomo Foundation and also by JSPS/MEXT KAKENHI (18687010, 21770109, and 23770110).

References (50)

  • J.K. Forwood et al.

    Quantitative structural analysis of importin-beta flexibility: paradigm for solenoid protein structures

    Structure

    (2010)
  • N. Fukuhara et al.

    Conformational variability of nucleo-cytoplasmic transport factors

    J. Biol. Chem.

    (2004)
  • A. Cook et al.

    The structure of the nuclear export receptor Cse1 in its cytosolic state reveals a closed conformation incompatible with cargo binding

    Mol. Cell

    (2005)
  • E. Conti et al.

    Karyopherin flexibility in nucleocytoplasmic transport

    Curr. Opin. Struct. Biol.

    (2006)
  • G.D. Van Duyne et al.

    Atomic structures of the human immunophilin FKBP-12 complexes with FK506 and rapamycin

    J. Mol. Biol.

    (1993)
  • E.A. Merritt et al.

    Raster3D: photorealistic molecular graphics

    Methods Enzymol.

    (1997)
  • D. Gorlich et al.

    Transport between the cell nucleus and the cytoplasm

    Annu. Rev. Cell Dev. Biol.

    (1999)
  • I.R. Vetter et al.

    The guanine nucleotide-binding switch in three dimensions

    Science

    (2001)
  • A. Cook et al.

    Structural biology of nucleocytoplasmic transport

    Annu. Rev. Biochem.

    (2007)
  • C. Klebe et al.

    Interaction of the nuclear GTP-binding protein Ran with its regulatory proteins RCC1 and RanGAP1

    Biochemistry

    (1995)
  • F.R. Bischoff et al.

    Catalysis of guanine nucleotide exchange on Ran by the mitotic regulator RCC1

    Nature

    (1991)
  • D. Gorlich et al.

    Identification of different roles for RanGDP and RanGTP in nuclear protein import

    EMBO J.

    (1996)
  • M. Fukuda et al.

    CRM1 is responsible for intracellular transport mediated by the nuclear export signal

    Nature

    (1997)
  • B. Ossareh-Nazari et al.

    Evidence for a role of CRM1 in signal-mediated nuclear protein export

    Science

    (1997)
  • T. la Cour et al.

    Analysis and prediction of leucine-rich nuclear export signals

    Protein Eng. Des. Sel.

    (2004)
  • Cited by (25)

    • Mechanistic Insights from Structural Analyses of Ran-GTPase-Driven Nuclear Export of Proteins and RNAs

      2016, Journal of Molecular Biology
      Citation Excerpt :

      Furthermore, the structure of free Xpo1p revealed that the C-terminal α-helix (referred to as the C-helix) adopts dramatically different conformation compared to the Xpo1p-cargo-Gsp1p complex, and it lies across the central cavity of the Xpo1p ring with its C-terminal tail interacting with the HEAT9 loop and also with the inner surface behind the NES-binding site. The C-terminal tail in free Xpo1p interacts with the HEAT9 loop in such a way that stabilizes the autoinhibitory conformation of the HEAT9 loop and thereby reinforces autoinhibiton [112]. The role of the C-terminal tail of CRM1 (Xpo1p) in autoinhibition has been supported by biochemical assays [112,113].

    • Structural insights into how Yrb2p accelerates the assembly of the Xpo1p nuclear export complex

      2014, Cell Reports
      Citation Excerpt :

      NES binds to the hydrophobic cleft on the outer surface of Xpo1p (CRM1), formed between the A helices of HEAT repeats 11 and 12, whereas Gsp1p-GTP (Ran-GTP) binds to the interior surface of Xpo1p (CRM1), making intimate contacts with HEAT repeats 1–4, 17, and 19 and a long β hairpin loop (referred to as HEAT9 loop) inserted between the A and B helices of HEAT repeat 9. HEAT9 loop plays a key role in the cooperative Gsp1p-GTP (Ran-GTP) and cargo binding to Xpo1p (CRM1) (Koyama and Matsuura, 2010; Saito and Matsuura, 2013). In free Xpo1p (CRM1), HEAT9 loop binds to the inner surface beneath the NES-binding cleft, stabilizing the cleft in a closed conformation that is incompatible with cargo binding.

    • Structure of a truncation mutant of the nuclear export factor CRM1 provides insights into the auto-inhibitory role of its C-terminal helix

      2013, Structure
      Citation Excerpt :

      The limited resolution of the diffraction data does not affect the validity of this analysis, because the latter is restricted to a discussion of conformational changes in which the shifts of helices or HEAT repeats exceed the positional uncertainty of these elements in our atomic model. The structure of CRM1ΔC is more similar to that of unbound cytosolic CRM1 (Monecke et al., 2013; Saito and Matsuura, 2012; root-mean-square deviation [rmsd] = 4.3 or 4.4 Å) than to the Ran- and cargo-bound nuclear conformation (Monecke et al., 2009; rmsd = 6.7 Å). In particular, the central region involved in NES recognition (including the HEAT-9 loop) is nearly identical to that of unbound CRM1 (aligning repeats 9–12 with unliganded yeast CRM1 yields an rmsd of 0.93 Å for 190 Cα atoms).

    View all citing articles on Scopus
    1

    Present address: N. Saito, Hamamatsu Photonics, Shizuoka, Japan.

    View full text